Biological Impacts of Hydraulic DisturbancesAssociated with Jet Boating on the Chilkat

River, Alaska

D.F. Hill1, B.D. YounkinDepartment of Civil & Environmental Engineering

The Pennsylvania State UniversityUniversity Park, PA 16802

January 2005

1Contact: 814.863.7305, dfh@engr.psu.edu

Contents

1 Introduction 111.1 Historical Overview of the Chilkat River Basin . . . . . . . . . 111.2 Review of Findings from the 2002 Study . . . . . . . . . . . . 121.3 Study Scope and Limitations . . . . . . . . . . . . . . . . . . . 131.4 Report Outline . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2 Review 142.1 Hydrodynamic Imapcts . . . . . . . . . . . . . . . . . . . . . . 142.2 Turbidity Impacts . . . . . . . . . . . . . . . . . . . . . . . . . 17

3 Field Study 193.1 Erosion Pin Sites and Data . . . . . . . . . . . . . . . . . . . 19

3.1.1 Second Year Erosion Data . . . . . . . . . . . . . . . . 283.2 Near-Bank Hydraulics . . . . . . . . . . . . . . . . . . . . . . 30

3.2.1 Velocity Data . . . . . . . . . . . . . . . . . . . . . . . 303.2.2 Turbidity Data . . . . . . . . . . . . . . . . . . . . . . 413.2.3 Near-Bank Dye Study . . . . . . . . . . . . . . . . . . 42

4 Laboratory Study 494.1 Experimental Facilities . . . . . . . . . . . . . . . . . . . . . . 49

4.1.1 Turbulence Trial . . . . . . . . . . . . . . . . . . . . . 504.1.2 Turbidity Trial . . . . . . . . . . . . . . . . . . . . . . 51

4.2 Experimental Procedures . . . . . . . . . . . . . . . . . . . . . 514.2.1 General Details . . . . . . . . . . . . . . . . . . . . . . 514.2.2 Turbulence Trials . . . . . . . . . . . . . . . . . . . . . 524.2.3 Turbidity Trials . . . . . . . . . . . . . . . . . . . . . . 53

4.3 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . 604.3.1 Turbulence Trials . . . . . . . . . . . . . . . . . . . . . 60

1

4.3.2 Turbidity Trials . . . . . . . . . . . . . . . . . . . . . . 62

5 Discussion and Conclusions 645.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 655.3 Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . 66

2

List of Figures

2.1 Schematic of rotational motion imparted by the spatial varia-tion of mean velocity. . . . . . . . . . . . . . . . . . . . . . . . 15

2.2 Schematic of a turbulent velocity signal. . . . . . . . . . . . . 16

3.1 Topographic view of the Chilkat River basin. Source: USGS7.5 minute Skagway B-3 & C-3 quadrangles. . . . . . . . . . . 21

3.2 Topographic closeup view of region (a) from Fig. 3.1. Source:USGS 7.5 minute Skagway B-3 & C-3 quadrangles. . . . . . . 22

3.3 Topographic closeup view of region (b) from Fig. 3.1. Source:USGS 7.5 minute Skagway B-3 & C-3 quadrangles. . . . . . . 23

3.4 Topographic closeup view of region (c) from Fig. 3.1. Source:USGS 7.5 minute Skagway B-3 & C-3 quadrangles. . . . . . . 24

3.5 Aerial photograph mosaic indicating several reference loca-tions in the Upper Chilkat Basin. . . . . . . . . . . . . . . . . 25

3.6 Sontek ADV in ‘side-looking’ mode. . . . . . . . . . . . . . . . 303.7 Sontek ADV in ‘downward-looking’ mode. . . . . . . . . . . . 313.8 Schematic of near-bank bathymetry. Note that the positive x

axis is pointing out of the page. . . . . . . . . . . . . . . . . . 323.9 Near-bank velocity data from the Disney Channel east bank

site. View is downstream. (a) - contours of mean streamwisevelocity u in units of cm s−1. (b) - contours of turbulent kineticenergy k in units of cm2 s−2. . . . . . . . . . . . . . . . . . . . 35

3.10 Near-bank velocity data from the Disney Channel west banksite. View is downstream. (a) - contours of mean streamwisevelocity u in units of cm s−1. (b) - contours of turbulent kineticenergy k in units of cm2 s−2. . . . . . . . . . . . . . . . . . . . 39

3

3.11 Near-bank velocity data from the Dry Lake site. View is down-stream. (a) - contours of mean streamwise velocity u in unitsof cm s−1. (b) - contours of turbulent kinetic energy k in unitsof cm2 s−2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

3.12 Top-view schematic of a steady-input, bank-discharge dye study.Note the widening of the dye plume with downstream distance. 43

3.13 Contours (in ppb) of dye concentration for the Dry Lake near-bank steady-discharge dye study. The symbols denote thelocations where the dye concentration was measured. . . . . . 45

3.14 Lateral profiles of dye concentration at various x locations. . . 463.15 Theoretical predictions for dye concentration at the Dry Lake

site, using the best-fit values for flow velocity and dispersioncoefficient. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

4.1 Photograph of the experimental arrangement of tanks. . . . . 564.2 Schematic sketch of an axisymmetric fluid jet. . . . . . . . . . 574.3 Schematic close-up views of the experimental apparatus for

the turbulence trials. . . . . . . . . . . . . . . . . . . . . . . . 584.4 Typical photograph of (sedated) fish being batch weighed. . . 59

4

List of Tables

3.1 Erosion pin data for the Upper Chilkat River. Values givenreflect the amount of erosion observed during the 12 monthperiod between summer 2002 and summer 2003. . . . . . . . . 26

3.2 Erosion pin data for the Upper Chilkat River. Values givenreflect the amount of erosion observed during the 24 monthperiod between summer 2002 and summer 2004. . . . . . . . . 29

3.3 Summary of velocity and turbidity data for the Disney Chan-nel (east bank) site. . . . . . . . . . . . . . . . . . . . . . . . . 33

3.4 Summary of velocity and turbidity data for the Disney Chan-nel (west bank) site. . . . . . . . . . . . . . . . . . . . . . . . 38

3.5 Summary of velocity data for the Dry Lake site. . . . . . . . . 403.6 Turbidity data, as obtained from channel ‘transects.’ . . . . . 423.7 Rhodamine dye concentrations observed during steady-input,

bank-discharge dye study at Dry Lake. . . . . . . . . . . . . . 47

4.1 Circulation loop volumetric flow rates, jet source velocities,and energy dissipation levels for the three treatment levels inthe turbulence trials. . . . . . . . . . . . . . . . . . . . . . . . 52

4.2 Target turbidity levels for the three treatments. . . . . . . . . 544.3 Summary of weight data for the turbulence trial. . . . . . . . . 614.4 Summary of weight data for the turbidity trial. . . . . . . . . 63

5

Conversion Factors

Every effort is made in this report to use the SI (metric) system of units. Thefollowing table of conversions will assist those wishing to convert reportedvalues in SI units to BG units.

Quantity Multiply by To Obtain

Area square meter 10.77 square footArea square kilometer 0.386 square mileArea square kilometer 247.1 acreEnergy joules 0.738 foot poundsFlow Rate cubic meters per second 35.31 cubic feet per secondFlow Rate cubic meters per second 15,850 gallons per minuteLength meter 3.281 footLength kilometer 0.622 mileMass kilograms 0.685 slugsMass kilograms 2.21 pounds massPower watts 0.00134 horsepowerPower watts 0.738 foot pounds per secondSpeed meters per second 2.237 miles per hourStress / Pascals 0.000145 pounds per square inchPressure

6

Nomenclature

The following lists and defines the variables that are used herein. Note that,in a few cases, a single symbol is used to denote multiple quantities. Everyeffort is made to clarify this in the text of the report.

Symbol Description

A areaC concentrationd50 median grain size diameterΦ dissipation of energyg gravitational accelerationγ fluid specific weighth depthhL head lossk turbulent kinetic energy

KL loss coefficientKy diffusion coefficientm mass flow rateµ absolute viscosityP powerQ dischargeρ fluid densityτ0 bank / bed shear stress

(u, v, w) velocity componentsu∗ shear velocityV velocityV mean velocityV ′ fluctuating velocityV volume

(x, y, z) spatial coordinates

7

Preface

The work described in this report was conducted for the Alaska Departmentof Fish and Game (ADFG), Sport Fish Division.

The field work was performed during June of 2003 by researchers fromthe Department of Civil and Environmental Engineering at the PennsylvaniaState University. Assistance with logistical details was provided by membersof ADFG. The laboratory work was performed during the winter and springof 2004 by researchers from the Department of Civil and EnvironmentalEngineering and the School of Forest Resources at the Pennsylvania StateUniversity. Assistance with the laboratory experiments was provided by TimStecko, Doug Fischer, and Justin Koller.

The authors wish to thank ADFG for providing many of the on-site re-sources required for the field study.

8

Summary

A two-part study has been conducted in an effort to understand the impacts,if any, that commercial jet boating is having on rearing juvenile salmonids inthe upper Chilkat River, Alaska. In the first part of the study, measurementsof near-bank fluid velocity and turbidity were made. The measurementswere made at a variety of depths and distances from the bank in an effortto build up a complete picture of the hydraulic climate in the near-bankregion. The goal of this component of the study was to determine the ambientcharacteristics typical of rearing habitat.

In the second part of the study, controlled laboratory experiments in-vestigating the biological response of salmon fry to varying levels of thesehydraulic disturbances were conducted. The goal of this component of thestudy was to determine whether or not salmon fry respond negatively to theelevated hydraulic disturbances that are associated with boat wakes breakingon the banks of the river.

Some of the key findings of this study include:

• Ongoing erosion pin measurements in the Chilkat River (since summer2002) demonstrate steady erosion of the banks in response the theloadings on the banks. These loadings include the stream flow andthe impacts of boat wakes. For the most part, larger erosion valuesare found in larger channels, but there are some exceptions to thisgeneralization.

• Numerous estimation methods applied to the field data arrived at ashear velocity of 5 − 10 cm s−1 in the vicinity of the banks. Thisparameter is of broad use in terms of characterizing the mean andturbulent flow fields.

• Field measurements consistently placed the ambient, or background,turbidity of the river at ∼ 100 NTUs (nephelometric turbidity units).

• Laboratory trials revealed that the presence of fluid turbulence reducedthe growth rate of salmon fry when compared to a control group ex-posed to no turbulence. However, the reduction in growth rate ap-peared not to be a function of the turbulence level in the tank.

• Laboratory trials revealed that the presence of turbidity did not seemto affect the growth rate of salmon fry. Moreover, the growth rate did

9

not appear to be function of the turbidity level in the tank.

10

Chapter 1

Introduction

1.1 Historical Overview of the Chilkat River

Basin

The Chilkat River basin, consisting of the Chilkat, Klehini, Tsirku, and Kel-sall Rivers, lies approximately 80 miles northwest of Juneau, Alaska, anddrains into the upper Lynn Canal. A significant portion of the basin over-laps with the 49,000 acre Chilkat Bald Eagle Preserve (CBEP). The preservewas designated by the State Legislature in 1982. Prior to this, the importanceof the large fall / winter gatherings of bald eagles was recognized through thedesignation of a 4800 acre Critical Habitat Area in 1973 and the adoption ofthe Haines / Skagway Land Use Plan in 1979.

As described in Alaska statute, and as reviewed by the Alaska Departmentof Natural Resources (ADNR, 2002), the CBEP was created to:

• Protect and perpetuate the Chilkat bald eagles and their essential habi-tats within the preserve,

• Protect and sustain the natural salmon spawning and rearing areas ofthe Chilkat River and Chilkoot River systems in perpetuity,

• Provide continued opportunities for research, study, and enjoyment ofbald eagles and other wildlife,

• Maintain water quality and necessary water quality,

11

• Provide for the continued traditional and natural resource based lifestyleof the people living in the area, and

• Provide for other public uses consistent with the primary purpose ofthe Preserve.

In 1985, a plan for managing the Preserve was adopted. During the periodof 2001-2002, a revision process for the plan was undertaken. One of the issuesof concern, and the one pertinent to the present study and report, is that thePreserve has experienced greatly increasing levels of commercial recreationaluse over the past 10 years. The specific commercial use of relevance to thepresent study is that of large-scale jet-boat tours of the upper Chilkat River,between Wells Bridge and the confluence with the Kelsall River.

1.2 Review of Findings from the 2002 Study

Concern that operation of these large jet-boats was having an impact onsalmon habitat prompted the ADFG to sponsor a field project in the sum-mer of 2002 (Hill et al., 2002). The main objective of that study was tomeasure the near-bank hydraulic parameters associated with jet-boating ac-tivity. These parameters included wake heights and turbidity levels.

The results of that study indicated that the boat wakes generated bycommercial traffic were capable of dislodging sediment from the banks of theriver. Peak values of recorded near-bank turbidity were found to far outweigh(by an order of magnitude) the ambient load of the river and were found tobe an increasing function of wake height.

Comparisons of estimates of parameters such as shear stress, energy, andpower suggested that, in some channels, the loading associated with boatwakes could exceed that of the streamflow. An equation relating wake heightto ‘sailing line’ distance was developed in an effort to provide guidelines forminimizing the impacts of boat wakes on the river banks.

In conclusion, the 2002 field study demonstrated that, despite the rela-tively low usage levels on the Chilkat River, the potential impacts associatedwith boating could not be ruled out.

12

1.3 Study Scope and Limitations

The goal of the present field and laboratory study has been to investigate theconsequences, if any, of these hydraulic loadings on juvenile salmon. ADFGbiologists have documented near-bank areas as being productive rearing habi-tat. These areas, where wakes shoal, break, and dislodge sediment are thosemost heavily impacted by boating activity.

The specific goals of the field study were to (i) revisit numerous sites whereerosion pins were placed during the 2002 field season and to (ii) obtain dataon the distribution of velocity (both mean and turbulent) and turbidity innumerous near-bank areas.

The specific goals of the laboratory study were to determine how juvenilecoho salmon respond to increasing levels of turbulence and turbidity. Whileit is of course impossible to replicate the exact field situation in a laboratorysetting, efforts have been made to reproduce certain elements, such as theintermittent and relatively short-duration nature of the loading associatedwith boating activity. In addition, native sediments have been brought backfrom the Chilkat River in order to provide the turbidity for the present trials.To gauge whether or not any systematic trend exists, a control group andthree treatment groups (each at a different loading level) were used for thelaboratory study.

1.4 Report Outline

A brief review of near-bank hydraulics, along with an extensive review ofthe biological response of fish to hydraulic stresses, is presented in Chapter2. Chapter 3 describes the field experiments and Chapter 4 describes thelaboratory experiments. Finally, Chapter 5 provides discussion of the results,conclusions, and a discussion of any additional work that should be done.

13

Chapter 2

Review

In this chapter, previous studies dealing with the biological impacts of tur-bulence and turbidity on juvenile fish are breifly reviewed.

2.1 Hydrodynamic Imapcts

First of all, it is worth discussing what exactly constitutes a hydrodynamicimpact. From a lay point of view, there is certainly the notion that impact isin some way associated with, and proportional to, the flow of water. However,it is not quite so simple. Fluid velocity, in and of itself, is not an indicator ofa stressful environment. In the case of a fluid flow that is spatially uniformand temporally steady, there will be no sensation of motion at all.

There are two basic mechanisms that can lead to a fluid flow exertinga ‘load’ on an object that is being carried along by the flow. First of all,and as illustrated in Fig. 2.1, in a flow that is steady in time, but that hasspatial variations in velocity, there are gradients in velocity present. Thesegradients are significant in that, for an object of some finite size, the fluidvelocity along one edge of the object may differ from that along anotheredge. These different velocities result in a shear stress which, in turn, seeksto impart rotational motion to the object. This rotational motion can leadto accelerations of a damaging magnitude.

The other basic mechanism, as illustrated in Fig. 2.2, is associated withthe temporal variations of a fluid flow and is therefore, by definition, associ-ated with the turbulent nature of a fluid flow. Here, although the velocitymay not vary significantly between closely adjoining points in space (the ve-

14

y

V

Figure 2.1: Schematic of rotational motion imparted by the spatial variationof mean velocity.

locity gradients are small), the velocity can, at a single point, fluctuate wildlyin time. As shown in the figure, the velocity V (t) recorded at some point inspace is a function of time. By sampling over a long interval, T , it is possibleto compute the time-averaged, or mean velocity;

V =1

T

∫ t=T

t=0

V dt.

Thus, at any instant in time, the velocity is given by the sum of the meanvelocity and the ‘fluctuating’ velocity, i.e. V (t) = V + V ′.

Now, the time-average of V ′ itself is, by definition, equal to zero and,therefore, not very useful. The root mean square value of the fluctuation,however, given by

V ′rms =

(1

T

∫ t=T

t=0

V ′2 dt

)1/2

,

is, in general, non-zero and a very useful measure of how turbulent the flowis. In particular, the ratio of the rms value of a velocity component to itsmean value is referred to as the turbulence intensity. A flow that has a highturbulence intensity experiences significant accelerations that, as with thecase of accelerations due to shear-induced rotation, can be damaging.

Morgan et al. (1976) studied the effects of shear stress on fish eggs andlarvae, specifically bass and perch. They used a system that consisted of

15

V

t

V

V'

Figure 2.2: Schematic of a turbulent velocity signal.

two concentric cylinders, one of which was rotating, the other of which wasfixed. In principle, the shear stress inside the annulus between the cylindersis a function of the cylinder diameters, rotation rate, and fluid properties,and can therefore be systematically varied. Batches of eggs were subjectedto long-term (2-3 day) exposure and a regression equation between mortal-ity and shear stress was developed. While of value in terms of its effort tosystematically investigate these effects in the laboratory, the authors inap-propriately applied a laminar theory to the case of a turbulent flow.

Killgore et al. (1987) studied the effects of turbulence on the yolk-saclarvae of paddlefish. This was accomplished by recirculating water, at eitherlow or high rates, through circular raceways. This hydrodynamic ‘loading’was done in an intermittent fashion. Unfortunately, this study also hadsome problems, from a fluid mechanical point of view. While the authorsrepeatedly referred to levels of ‘turbulence,’ their manuscript gave levels ofshear stress. Therefore, it is not entirely clear what parameter was beingsystematically varied.

Sutherland & Ogle (1975) conducted a lab study on the effects of jet boatson salmon eggs. Initially, field experiments were performed by operating jetboat over pressure sensors buried in gravel beds. The observed pressuregradients were later reproduced in a laboratory setting where gravel andsalmon eggs were packed into a vertical column. The eggs were exposed tothe loading for a time period on the order of one minute. After an individual

16

test, the fatality rate was measured. Tests were performed on eggs of different‘ages,’ from 6 to 13 days and fatalities averaged about 40 percent.

Much more recently, a report by Neitzel et al. (1998) discusses the expo-sure of small (5-10 cm) fish to extremely high-level, short-duration bursts ofshear stress. The main motivation for this work had to do with fish passagethrough hydroelectric turbines. The high-stress environment was createdwith a high-velocity water jet, discharging into an ambient body of water.The fish were injected into the ‘edge’ of this jet through a small feeder tube.The biological stress on the fish was measured by subsequently inspectingthe fish for death and significant injury. Many different species (e.g. salmon,trout, shad, etc.) were studied and it was found that there was not verymuch variation in response among the species.

Finally, a study by Rehmann et al. (2003) studied the effects of fluidturbulence on the mortality of zebra mussel larvae. From a fluid mechanicalpoint of view, this study was very well done. An acoustic doppler velocimeterwas used to characterize the velocity field induced by a bubble plume in atank. The frequency spectra of the time series of the velocity data werethen fit to the theoretical Batchelor spectrum allowing for the estimationof the dissipation of turbulent kinetic energy. This, in turn, allowed forthe estimation of the Kolmogorov scale, which is the scale of the smallestturbulent eddies in the flow. It was found that the greatest mortality occurredwhen the larvae were approximately equal in size to this scale.

2.2 Turbidity Impacts

Gregg & Bergersen (1980) studied the effects of turbidity (and turbulence,incidentally) on Mysis relicta (opossum shrimp). Three turbidity and fourturbulence levels were considered, for a total of 12 trials. As with some of thepreviously mentioned studies, the way in which the authors defined ‘turbu-lence’ was unsound. After initiating the experiment, mysids were examinedon days 2, 4, 6, and 8 in order to determine mortality. The results indicatedthat turbidity was not that important in contributing to mortality.

Morgan et al. (1983) considered the effects of sediment on the eggs andlarvae of bass and perch. The study found that percent hatch was not sig-nificantly affected by suspended sediment concentration levels up to about 5grams per liter, but that developmental rates were severely lowered at con-centrations above 1.5 grams per liter. Larvae showed significant mortality

17

when exposed to 24 and 48 hour durations of high turbidity. Arruda et al.(1983) studied the effects of turbidity on zooplankton and showed that mod-est levels severely reduced feeding rates. Finally, Breitburg (1988) studiedthe effects of turbidity on prey consumption by striped bass larvae. Thesewere very short-duration (half hour) experiments. Tanks of different turbid-ity levels were set up and it was found that high turbidity levels reduced thenumber of copepods captured by the larvae.

Turning attention to the effects of turbidity on fish at a more advanceddevelopmental stage, an excellent review is provided by Bash et al. (2001).First, the authors discuss how effects can be categorized as lethal, sub-lethal,and behavioral. Also, they discuss how duration of exposure, frequency ofexposure, level of exposure, and life stage of the fish all play a significant role.They review laboratory studies that detailed gill trauma, elevated bloodsugars, and elevated plasma glucose (Servizi & Martens, 1987), as well asreduced growth rates (Sigler et al., 1984). Regarding the growth rates, thework by Sigler et al. (1984) showed that turbidities as low as 25 NTUs (NTU- nephelometric turbidity unit) reduced growth.

Moving on to behavioral effects, Sigler et al. (1984) showed that fishwould migrate away from regions of relatively high turbidity (∼ 150 NTUs)to regions of relatively low turbidity (∼ 60 NTUs). Newly emerged fry wereshown to be even more susceptible. Berg (1982) showed that streams affectedby frequent short-term sediment pulses with concomitant loss of territorialitymay experience a decrease in growth and feeding rates. Berg & Northcote(1985) showed that, during turbid phases (∼ 30−60 NTUs), territories brokedown and subordinate fish captured a greater proportion of the prey. Otherstudies (Berg, 1982) have shown that fairly low turbidity levels (∼ 20 − 60NTUs) caused fish to mis-strike prey items. Another issue (Newcombe &MacDonald, 1991) is that increased turbidity levels may affect the benthiccommunity and therefore decrease the food supply.

Bash et al. (2001) also reviewed the current state and provincial tur-bidity standards for the Pacific Northwest, British Columbia, and Alaska.For Alaska, the standard is 25 NTUs above natural condition. Note thatthis standard is an ‘end-of-pipe’ standard, unless a mixing zone has beenapproved.

18

Chapter 3

Field Study

The Chilkat River Basin is illustrated in Figs. 3.1-3.4. Basic hydrologicinformation on the basin is provided by Bugliosi (1985) and is reviewed byHill et al. (2002). The work in both the 2002 and the 2003 field seasons tookplace in the upper basin, above Wells Bridge. Figure 3.5 presents a mosaic ofaerial photographs of the upper basin. In addition, key reference locations arenoted on this figure. Throughout much of the rest of this chapter, referencewill be made to many of the sites listed.

3.1 Erosion Pin Sites and Data

An erosion pin is a simple, yet effective, way to measure the amount oferosion that occurs at a given location. A pin is nothing more than a thinand long metal rod that is driven perpendicularly into the surface (river bank,river bed, etc.) of interest. By measuring, at an initial time and then somesubsequent time, the length of the pin that is exposed, the amount of erosionthat occurred between the two times is immediately determined.

During the 2002 field experiment, pins were placed at a number of lo-cations. At each location, three pins were placed. All pins were ∼ 1 cmdiameter rebar and were driven approximately 1 m into the bank, leavingapproximately 15 cm initially exposed. Of the three pins, one was placed atthe current level of the water, the second was placed 15 cm vertically above,and the third was placed 15 cm vertically below. All pins were placed inJune. The pin locations, including GPS waypoints where available, are givenby:

19

1. LC1 (N 59.29.269 W 136.03.652) - This lies above the Kelsall boatlaunch in the ‘left’ (looking upstream) channel. Pins were placed inboth the west and east banks of the channel.

2. LC2 (N 59.29.006 W 136.03.208) - This lies approximately halfwaybetween LC1 and the boat launch. Pins were placed in both the westand east banks of the channel. The west pins are slightly upstream ofa large culvert.

3. MC1 - This is a site near the southern end of what is referred to asDry Lake. This is in the ‘middle’ channel above the boat launch. Pinswere placed in the west bank only.

4. SHEEP CYN (N 59.27.948 W 135.59.245) - This location is aboutthree fourths of the way up the small clearwater channel leading intoSheep Canyon Lake. Pins were placed in both the west and east banksof the channel.

5. EAST CHAN (N 59.25.605 W 135.55.828) - This is in the east chan-nel, not too far above Wells Bridge. Pins were placed in both the westand east banks.

6. EC 2 - This location is in the east channel, in between EAST CHANand SHEEP CYN. Pins were placed at two locations in the west bankonly. The lower set is slightly above the ‘tip’ of the island; the upperset is approximately 30 meters north of the lower set.

7. MC3 EAST (N. 59.29.025 W 136.02.932) This location is approxi-mately halfway between the Kelsall boat launch and Dry Lake. It isjust above a ‘point’ separating this middle channel from a very smalltributary on the right.

8. MC3 WEST (N 59.28.990 W 136.02.986) - This location lies more orless directly across the channel from MC3 EAST.

9. KELSALL - This location is on the west bank of the main chan-nel, slightly upstream of what is referred to as Jacquot’s Landing andslightly downstream of the confluence of the Chilkat and Kelsall Rivers.

During the 2003 field experiment, all pin locations were revisited and theerosion amounts were recorded. These erosion amounts are summarized inTable 3.1.

20

Klehini River

Chilkat River

Chilkat Lake

Wells Bridge

Klukwan

Tsirku River Chilkat River

(a)

(b)

(c)

Figure 3.1: Topographic view of the Chilkat River basin. Source: USGS 7.5minute Skagway B-3 & C-3 quadrangles.

21

Chilkat River

Kelsall River

Figure 3.2: Topographic closeup view of region (a) from Fig. 3.1. Source:USGS 7.5 minute Skagway B-3 & C-3 quadrangles.

22

Klukwan

Wells BridgeKlehini River

Tsirku River

Chilkat River

Figure 3.3: Topographic closeup view of region (b) from Fig. 3.1. Source:USGS 7.5 minute Skagway B-3 & C-3 quadrangles.

23

Chilkat RiverTsirku River

Klukwan

Chilkat Lake

Figure 3.4: Topographic closeup view of region (c) from Fig. 3.1. Source:USGS 7.5 minute Skagway B-3 & C-3 quadrangles.

24

Kelsall Confluence

Dry Lake

Kelsall Landing

Sheep Canyon

East Channel

Wells Bridge

DisneyChannel

Figure 3.5: Aerial photograph mosaic indicating several reference locationsin the Upper Chilkat Basin.

25

Loca

tion

Not

esU

pper

Pin

(cm

)M

iddle

Pin

(cm

)Low

erP

in(c

m)

Ave

rage

(cm

)

LC

1W

est

ban

k2.

50

00.

8E

ast

ban

k2.

57.

52.

54.

2LC

2W

est

ban

k0

00

0E

ast

ban

k0

00

0M

C1

35.5

35.5

35.5

35.5

SH

EE

PC

YN

Wes

tban

k20

.312

.710

.214

.4E

ast

ban

k0

7.6

7.6

5.1

EA

ST

CH

AN

Wes

tban

k10

.22.

52.

55.

1E

ast

ban

k0

02.

50.

8M

C3

EA

ST

35.5

35.5

35.5

35.5

MC

3W

EST

30.4

45.7

27.9

34.7

EC

2U

pst

ream

loca

tion

5.1

5.1

5.1

5.1

EC

2D

ownst

ream

loca

tion

7.6

7.6

05.

1K

ELSA

LL

Mis

sing1

Mis

sing

Mis

sing

n/a

Tab

le3.

1:E

rosi

onpin

dat

afo

rth

eU

pper

Chilka

tR

iver

.Val

ues

give

nre

flec

tth

eam

ount

ofer

osio

nob

serv

eddur-

ing

the

12m

onth

per

iod

bet

wee

nsu

mm

er20

02an

dsu

m-

mer

2003

.

1O

nepi

nw

asfo

und

layi

ngon

the

bank

near

by.

26

The relative erosion amounts given in the table are, for the most part,fairly intuitive. For example, the erosion observed in the ‘east channel,’ andthe ‘left channel’ (above the Kelsall boat launch) ranged from minimal to10 cm. Note that boating traffic in the left channel is minimal. During the2002 field study, the flow in the east channel was found to be on the orderof 15 m3 s−1. During the 2003 field study, the flow in the left channel wasmeasured with a wading survey and was found to be roughly 17 m3 s−1.

The erosion observed in the middle channel above the Kelsall landing wasmuch greater, averaging 35 cm. The flow in this channel was measured in2002 to be 73 m3 s−1 and in 2003 to be 67 m3 s−1. There are numerousreasons for these relatively large erosion values. First of all, the greater riverdischarge is expected to result in greater erosion. Second of all, the wakeheights recorded in these areas during the 2002 study were quite high, in theneighborhood of 25 cm. Third, the banks where these pins were installedwere fairly ‘raw,’ lacking substantial woody vegetation. The fairly exposednature of these banks also manifested itself in the turbidity measurementsof 2002, which showed that significant sediment sloughed off due to wakeimpact.

Most interesting, perhaps, are the values for erosion observed in the chan-nel leading to Sheep Canyon Lake. The discharge measurements from 2002revealed the flow in that channel to be exceedingly low, on the order of 1 m3

s−1. This, given the channel geometry, corresponds to an area-averaged flowvelocity of less than 10 cm s−1. Such a low velocity suggests that erosion dueto streamflow is negligible. While not definitive, this nevertheless points toboating activity as the agent causing erosion in this channel. Note that thisportion of the route run by the tour operator is an ‘out and back,’ with theramification that each boat travels the channel twice. The wake measure-ments from 2002 showed that the maximum wave heights were an extremelystrong function of boat speed. Planing speeds near 20 mph were capable ofgenerating wakes on the order of 30 cm. Extremely slow, ‘no-wake’ speeds,on the other hand, successfully reduced wake heights below 10 cm. It mustbe stressed that boat speed must be extremely slow (less than 5 mph) inorder to take advantage of this reduction; one trial with the boat travellingat 6 mph resulted in wake heights of ∼25 cm.

27

3.1.1 Second Year Erosion Data

During the summer of 2004, ADFG personnel returned to a limited number oferosion pin sites to record additional values. As the pins were not reset duringthe 2003 field study, the values tabulated in Table 3.2 reflect cumulativeerosion amounts for the 24 month period between summer 2002 and summer2004. As the data indicate, continued erosion was observed at all sites.

28

Loca

tion

Not

esU

pper

Pin

(cm

)M

iddle

Pin

(cm

)Low

erP

in(c

m)

Ave

rage

(cm

)

SH

EE

PC

YN

Wes

tban

k59

.735

.612

.736

.0E

ast

ban

k0

10.2

17.8

9.3

EA

ST

CH

AN

Wes

tban

k15

.212

.727

.918

.6E

ast

ban

k0

2.5

8.9

3.8

MC

3W

EST

44.5

49.5

36.8

43.6

EC

2D

ownst

ream

loca

tion

7.6

11.4

20.3

13.1

KE

LSA

LL

2.52

5.1

5.1

4.2

Tab

le3.

2:E

rosi

onpin

dat

afo

rth

eU

pper

Chilka

tR

iver

.Val

ues

give

nre

flec

tth

eam

ount

ofer

osio

nob

serv

eddur-

ing

the

24m

onth

per

iod

bet

wee

nsu

mm

er20

02an

dsu

m-

mer

2004

.

2R

ecal

ltha

tal

lthr

eepi

nsw

ere

mis

sing

duri

ngth

e20

03su

rvey

.N

ewpi

nsw

ere

plac

edat

that

tim

ean

dth

ese

valu

esre

flect

the

12m

onth

peri

odbe

twee

nsu

mm

er20

03an

dsu

mm

er20

04.

29

3.2 Near-Bank Hydraulics

The wading surveys of the 2002 field study give information on the averageflow properties in the various channels studied, but yield little in terms of thedetails of the flow near the banks. One of the objectives of the 2003 field studywas to take a closer look at the structure of the mean and fluctuating velocityfields in the near-bank regions. Of additional interest was the structure ofthe turbidity in the near-bank regions.

3.2.1 Velocity Data

To measure the velocity, use was made of a SonTek 10 MHz acoustic dopplervelocimeter (ADV), shown in Fig. 3.6. The ADV is capable of measuringthree components of velocity (u, v, w) at sampling rates of up to 20 hertz.Note that u denotes the component of velocity in the streamwise direction, vthe component in the lateral direction and w the component in the verticaldirection. The ADV operates by emitting acoustic energy from a transmitterlocated at the end of the probe. These signals are relected off of tiny particlesin the water and scattered back to three receivers. By analyzing the Dopplershift in the frequency, the velocity of the particles, and, by assumption, thecarrier water can be deduced.

Figure 3.6: Sontek ADV in ‘side-looking’ mode.

For the measurements described below, the ADV was used in its ‘downward-looking’ mode. The probe was mounted, as shown in Fig. 3.7, on a telescopingsurveying tripod, equipped with levels. In this fashion, the probe could beplaced at a prescribed horizontal and vertical location relative to the bank.Through a combination of vertical and horizontal traverses of the probe, arelatively detailed two-dimensional map of the velocity field can be built up.

30

Figure 3.7: Sontek ADV in ‘downward-looking’ mode.

A schematic configuration is shown in Fig. 3.8. As shown, a coordinatesystem is fixed at the waterline at the bank. In this case, the positive x axisis pointing upstream, the positive y axis is pointing across the river and thepositive z axis is pointing upwards.

Disney Channel East Bank

Velocity surveys were performed at several locations, the first of which wasalong the east bank of the Disney Channel, slightly downstream of whereit branches off from the main channel of the Chilkat River. The dischargein this channel could not be measured directly, but was estimated to be onthe order of 50 m3 s−1. The ADV was positioned at 23 different locationsand, at each location, a long time-series of three-dimensional velocity wasobtained. From each time series, and for each component, the mean velocityand the root-mean-square value of the fluctuating velocity was determined.

31

Looking Downstream

xy

z

Figure 3.8: Schematic of near-bank bathymetry. Note that the positive xaxis is pointing out of the page.

In addition, the Reynolds stresses were determined. These stresses are thetime-averages of the products between two fluctuating velocities, e.g. u′v′.Table 3.3 summarizes these data.

32

y(m

)z

(cm

)u

(cm

s−1)

v(c

ms−

1)

w(c

ms−

1)

√ u′2

(cm

s−1)

√ v′2

(cm

s−1)

√ w′2

(cm

s−1)

u′ v

′

(cm

2

s−2)

u′ w

′

(cm

2

s−2)

v′ w

′

(cm

2

s−2)

Turb

idity

(NT

Us)

1.55

-13.

2-2

5.7

2.55

-2.3

53.

233.

201.

88-0

.93

2.09

-0.8

51.

55-2

5.1

-24.

52.

64-3

.29

4.14

2.84

2.21

-2.4

14.

12-0

.26

141

1.55

-30.

4-2

4.6

3.22

-3.3

63.

383.

041.

871.

292.

53-0

.56

137

2.44

-13

-32.

86.

17-3

.05

4.01

3.08

2.23

-1.9

02.

09-1

.04

2.44

-33

-30.

32.

24-2

.48

4.34

3.16

2.83

5.01

4.04

1.31

133

2.44

-53.

1-2

7.9

3.13

-0.1

43.

813.

222.

61-1

.07

2.79

-1.1

513

42.

44-6

3.2

-22.

74.

351.

013.

543.

231.

95-2

.25

2.53

-2.6

113

43.

58-1

5.1

-52.

12.

73-5

.07

6.60

5.47

3.89

12.1

6.11

1.52

3.58

-35.

1-4

8.9

4.28

-4.5

57.

366.

495.

341.

8713

.11.

1814

03.

58-5

5.1

-43.

48.

65-3

.50

6.84

5.83

5.70

1.23

16.5

0.68

148

3.58

-75.

1-3

3.0

8.72

-1.2

86.

675.

394.

51-7

.84

9.63

-0.0

913

63.

58-9

0.6

-21.

97.

53-0

.37

5.61

4.79

2.23

-4.0

83.

57-1

.15

133

4.67

-15.

6-5

8.9

0.63

-5.7

76.

856.

094.

279.

707.

760.

044.

67-3

5.6

-60.

61.

79-6

.49

8.39

6.40

5.25

15.4

16.1

-1.8

714

14.

67-5

5.6

-49.

28.

07-4

.96

7.12

6.99

5.69

4.20

10.8

-12.

113

34.

67-7

5.6

-46.

54.

43-5

.51

8.24

7.25

6.09

10.7

14.5

-12.

313

14.

67-9

5.6

-35.

6-0

.87

-3.6

66.

726.

064.

958.

6712

.1-4

.59

126

4.67

-100

.8-3

2.9

-2.5

5-2

.46

7.69

6.97

4.28

20.7

4.66

-7.0

412

36.

17-1

7.9

-67.

69.

45-6

.47

6.76

5.59

4.13

5.19

3.11

-0.9

86.

17-3

7.9

-71.

49.

89-6

.25

7.07

5.70

4.69

5.67

2.83

-1.9

413

7Tab

le3.

3:Sum

mar

yof

velo

city

and

turb

idity

dat

afo

rth

eD

isney

Chan

nel

(eas

tban

k)

site

.

33

6.17

-57.

9-7

1.6

5.06

-4.5

07.

846.

365.

355.

0611

.7-1

.60

138

6.17

-77.

9-6

5.2

1.35

-2.5

08.

046.

325.

4113

.914

.2-2

.98

138

6.17

-97.

9-5

9.0

2.98

-0.5

49.

086.

484.

0312

.09.

37-4

.56

138

Tab

le3.

3:(c

onti

nued

)

34

To give a graphical feel to some of this data, consider first the streamwise(in the x direction) velocity u. Figure 3.9(a) shows a two-dimensional contourmap of u, which is the time-averaged, or mean, value. Figure 3.9(b) shows atwo-dimensional contour map of the turbulent kinetic energy, which is givenby

k =1

2(u′2 + v′2 + w′2). (3.1)

The structure of this mean flow is as expected, with boundary layer behaviorbeing observed both in the vertical and the lateral directions.

0 1 2 3 4 5 6

−1

−0.5

0

y (m)

z (m

)

−30−40

−50 −60

−70

0 1 2 3 4 5 6

−1

−0.5

0

y (m)

z (m

)

70

604020

(a)

(b)

Figure 3.9: Near-bank velocity data from the Disney Channel east bank site.View is downstream. (a) - contours of mean streamwise velocity u in unitsof cm s−1. (b) - contours of turbulent kinetic energy k in units of cm2 s−2.

The structure of the turbulent velocity is not as clearly evident, but meritssome discussion. For simplicity, consider for the moment flow in a widerectangular channel, so that the flow is essentially two-dimensional. In this

35

case, of course, the mean flow varies asymptotically from zero at the bed toa maximum at the free surface. Turbulent quantities, however, are foundto exhibit the opposite behavior with larger values near the boundary thannear the free surface. This is because the production of turbulence originatesfrom the shear at the boundary.

Nezu and Nakagawa (1993) give the following empirical formulae for thevertical distributions of turbulent quantities in an open channel flow:√

u′2

u∗= 2.30e−z/h (3.2)√

v′2

u∗= 1.63e−z/h (3.3)√

w′2

u∗= 1.27e−z/h (3.4)

k

u2∗

= 4.78e−2y/h. (3.5)

Note that the origin of the z coordinate now lies on the bed, h is the flowdepth, and u∗ is the shear velocity. This shear velocity is related to the shearstress on the bed, τ0, through the relationship u∗ =

√τ0/ρ, where ρ is the

density of the water.From these formulae, it can be shown that the depth-averaged value of√

u′2 is approximately 150% of the shear velocity. From the data, this yieldsa shear velocity of about 4.2 cm s−1 for the present field site. Performingsimilar estimates based upon the data and formulae for the vertical andlateral turbulent velocities yields estimates of 5.0 and 5.2 cm s−1. Thus, anoverall average of 4.8 is adopted for the shear velocity.

If, on the other hand, the formula for the turbulent kinetic energy isconsidered, depth-averaging and use of the data suggests a shear velocity of5.4 cm s−1. The point is that all of the turbulent velocity data provide fairlyconsistent estimates.

Disney Channel West Bank

This site was directly across the channel from the previous site. As before,the two-dimensional structure of the velocity in the near-bank area was de-termined and the data are summarized in the table below. The contours of

36

mean streamwise velocity and turbulent kinetic energy are given in Fig. 3.10.Following the same reasoning as discussed above, estimates of shear velocityare obtained from the root-mean-square values of the fluctuating velocitycomponents and from the turbulent kinetic energy. These varying estimatesgive, on average, a value of 4.8 cm s−1, similar to that found for the eastbank site.

37

y(m

)z

(cm

)u

(cm

s−1)

v(c

ms−

1)

w(c

ms−

1)

√ u′2

(cm

s−1)

√ v′2

(cm

s−1)

√ w′2

(cm

s−1)

u′ v

′

(cm

2

s−2)

u′ w

′

(cm

2

s−2)

v′ w

′

(cm

2

s−2)

Turb

idity

(NT

Us)

0.76

-21.

7-1

.12

0.92

0.80

2.20

1.91

1.29

-0.8

20.

520.

770.

76-2

6.8

-0.0

81.

340.

032.

982.

811.

37-1

.73

-0.4

40.

821.

52-1

7.0

-14.

50.

89-1

.88

5.70

3.69

2.61

-11.

1-0

.07

-0.4

01.

52-3

2.0

105

1.52

-47.

0-8

.08

-3.0

0-4

.49

4.76

4.03

2.94

-5.6

43.

862.

2695

1.52

-62.

0-1

.65

-3.8

7-4

.85

3.34

2.71

2.23

-2.6

01.

220.

9391

2.51

-9.6

-41.

815

.7-3

.69

6.71

5.81

3.95

-11.

8-7

.19

0.24

2.51

-29.

6-4

6.7

18.4

-4.3

56.

145.

794.

43-1

0.2

-3.5

3-0

.38

111

2.51

-49.

6-4

5.6

17.9

-4.4

48.

367.

816.

44-3

0.4

18.4

-10.

211

12.

51-6

9.6

-34.

011

.7-3

.65

8.60

8.47

6.50

-29.

318

.4-1

0.2

105

2.51

-89.

6-2

3.5

8.77

-5.0

57.

448.

175.

10-1

5.6

16.8

-0.5

496

3.40

-16.

2-5

6.5

24.5

-8.0

76.

106.

544.

40-1

2.3

-1.9

0-2

.56

3.40

-36.

2-6

3.5

27.9

-8.3

96.

255.

744.

63-7

.67

1.88

-2.4

911

73.

40-5

6.2

-69.

37.

01-7

.68

6.78

5.56

5.04

-3.3

86.

510.

6611

53.

40-7

6.2

-65.

75.

57-6

.35

8.72

6.85

5.97

-5.1

020

.32.

3912

53.

40-9

6.2

-50.

413

.8-5

.10

10.4

9.00

7.20

21.5

38.1

-2.3

013

53.

40-1

11.3

-34.

811

.8-4

.25

14.7

11.3

5.94

-46.

136

.1-6

.87

119

Tab

le3.

4:Sum

mar

yof

velo

city

and

turb

idity

dat

afo

rth

eD

isney

Chan

nel

(wes

tban

k)

site

.

38

−3.5 −3 −2.5 −2 −1.5 −1 −0.5 0

−1

−0.5

0

y (m)

z (m

)

−15

−25

−35−45

−55−65

−3.5 −3 −2.5 −2 −1.5 −1 −0.5 0

−1

−0.5

0

y (m)

z (m

)

100

9070

50 30

(a)

(b)

Figure 3.10: Near-bank velocity data from the Disney Channel west banksite. View is downstream. (a) - contours of mean streamwise velocity u inunits of cm s−1. (b) - contours of turbulent kinetic energy k in units of cm2

s−2.

Dry Lake

This site, which was one of the sites studied during the 2002 field project,lies on the west bank of the southern end of Dry Lake. The river herecarries a flow that is approximately 50% greater than that in the DisneyChannel. The obtained velocity data are presented in the Table below andare shown graphically in Fig. 3.11. If the data are considered along with theempirical formulae of Nezu and Nakagawa (1993), an average shear velocityof approximately 7.5 cm s−1 is found. This higher value is intuitive, giventhat the discharge, and hence shear stress, is greater at this site.

39

y(m

)z

(cm

)u

(cm

s−1)

v(c

ms−

1)

w(c

ms−

1)

√ u′2

(cm

s−1)

√ v′2

(cm

s−1)

√ w′2

(cm

s−1)

u′ v

′

(cm

2

s−2)

u′ w

′

(cm

2

s−2)

v′ w

′

(cm

2

s−2)

0.36

-10.

0-5

.41

-9.8

10.

437.

235.

695.

77-1

1.3

-2.8

1-2

.09

0.36

-20.

0-6

.34

0.63

2.45

7.29

5.88

6.35

-14.

93.

294.

430.

36-3

0.0

-8.8

33.

511.

477.

125.

686.

10-1

4.9

-1.6

32.

730.

36-3

5.0

-9.3

65.

301.

808.

655.

915.

85-1

6.5

-1.5

510

.60.

94-1

0.0

-49.

11.

11-2

.75

11.0

7.74

6.04

-26.

2-1

0.2

-2.6

50.

94-2

5.0

-57.

54.

86-3

.36

8.48

7.28

6.70

-15.

51.

79-3

.90

0.94

-40.

0-5

5.4

4.54

-3.5

49.

427.

406.

31- 18

.93

11.7

0.89

0.94

-55.

0-4

4.8

-5.6

4-2

.09

12.1

8.32

7.13

-14.

221

.211

.50.

94-7

0.0

-28.

9-6

.29

-1.5

811

.18.

716.

08-1

6.1

14.7

12.2

1.65

-10.

0-7

6.4

1.51

-8.0

69.

487.

895.

12-1

6.8

15.4

0.21

1.65

-25.

0-7

5.0

1.30

-8.8

110

.88.

596.

97-6

.428

.14.

111.

65-4

0.0

-65.

80.

24-5

.60

11.4

8.72

7.51

-0.8

635

.90.

111.

65-5

5.0

-59.

6-4

.86

-5.7

211

.78.

437.

018.

5338

.7-0

.48

1.65

-70.

0-4

7.8

-7.8

1-4

.23

12.1

9.59

7.44

6.08

43.4

3.34

1.65

-80.

0-3

6.8

-9.2

0-3

.60

10.8

10.2

6.05

8.33

26.5

4.85

Tab

le3.

5:Sum

mar

yof

velo

city

dat

afo

rth

eD

ryLak

esi

te.

40

−1.6 −1.4 −1.2 −1 −0.8 −0.6 −0.4 −0.2 0−1

−0.5

0

y (m)

z (m

)

−25

−45

−65

−1.6 −1.4 −1.2 −1 −0.8 −0.6 −0.4 −0.2 0−1

−0.5

0

y (m)

z (m

)

130

11585

70

(a)

(b)

Figure 3.11: Near-bank velocity data from the Dry Lake site. View is down-stream. (a) - contours of mean streamwise velocity u in units of cm s−1. (b)- contours of turbulent kinetic energy k in units of cm2 s−2.

3.2.2 Turbidity Data

The report from the 2002 field project extensively documented the increasesin near-bank turbidity that would occur in response to wakes of differing sizes.Elevated turbidities of two to five times the ambient value were commonlyobserved, with occasional amplifications of up to 10. During the 2003 fieldproject, a more extensive survey of the background turbidity in the river wasundertaken.

First, an OBS sensor was mounted in tandem with the ADV during thevelocity trials described above. In this fashion, just as the two-dimensionalstructure of the velocity could be evaluated, so could the structure of the tur-bidity. The observed turbidity values are given in the tables of the previous

41

section and indicate that there is little spatial variability. A typical value ofambient turbidity is ∼ 100 NTUs.

In addition to these measurements, representative values of ambient tur-bidity were obtained in a second manner. At a total of seven sites, rangingfrom Wells Bridge in the south to the Kelsall Confluence in the north, ‘tran-sects’ of turbidity were obtained. These were obtained by idling across thewidth of the river over a 2 minute period at each site. During the traverse,turbidity data were continuously being taken at a depth of ∼ 50 cm belowthe surface. Upon completion of the traverse, the data were averaged, inorder to arrive at a representative value. As Table 3.6 indicates, there was,again, little variability in the data, with all sites pointing to approximately100 NTUs.

Location Average Turbidity (NTUs)

Main channel, just above Disney Channel 118Disney Channel 93East Channel 115

Main Channel, above Kelsall Landing 105Kelsall confluence 91

Wells Bridge 120

Table 3.6: Turbidity data, as obtained from channel ‘transects.’

3.2.3 Near-Bank Dye Study

The level of the turbulence in an open channel flow can also be estimated byassessing the flow’s ability to mix a scalar. In laminar flows, a scalar suchas a passive dye will mix only by the relatively slow process of moleculardiffusion. In a turbulent flow, the dye is more efficiently mixed by turbulenteddies of varying length and time scales. For a thorough introduction to themechanics of river mixing, the reader is referred to Rutherford (1994).

In the interest of a brief introduction, consider the schematic overheadview of a river of rectangular cross section and large aspect ratio (i.e. width�depth) shown in Fig. 3.12. Next, consider the situation where dye, of someknown initial concentration, is steadily discharged at the bank at x = 0.As the flow carries the dye downstream, turbulent eddies will very rapidlymix the dye uniformly across the depth of the flow. At this point, the dye

42

concentration field becomes two-dimensional, exhibiting variation only in thex − y plane. With further downstream distance, the turbulent eddies willcontinue to mix the dye across the width of the river. As a consequence, theconcentrations observed at the bank will decrease with increasing x.

x

y

Input

Plume Boundary

Figure 3.12: Top-view schematic of a steady-input, bank-discharge dye study.Note the widening of the dye plume with downstream distance.

Experimental Instrumentation

For the present experiment, measurements of dye concentration were madewith a Turner 10-AU Field Fluorometer. The Turner Fluorometer has asensitivity of 10 parts per trillion of Rhodamine WT (which was the dyeused). The fluorometer was used in continuous sampling mode, which meansthat a small centrifugal pump was used to continuously pump river waterthrough the flow-cell of the fluorometer. An internal data logger was used torecord the time-series of fluorescence. Subsequent to the field measurements,the data were downloaded to a laptop for analysis.

Experimental Procedure

A steady-input bank-discharge dye study was conducted on the west bank ofDry Lake. This site was chosen as velocity data had previously been obtainedat the site. This would allow for a comparison / verification of the resultsobtained by the two methods.

The experiment was conducted by first preparing a dilute solution ofRhodamine WT dye. Rhodamine WT dye is commonly sold in solution with

43

20% active ingredient. As this solution is somewhat more dense than water, itis common to further dilute the solution with ambient water. For the presentexperiment, a solution was prepared that had a concentration of 5.1%, byvolume, of the initial 20% solution.

This solution was then discharged, at a rate of 48 ml min−1 at the bankof the river. After allowing some time for the dye plume to establish itself,measurements were made at a variety of x and y locations downstream of theinjection point. At each location, data were sampled continuously at a rateof 1 Hz for approximately 2 minutes. Each time-series was then averaged toobtain an average value of concentration at that point. The data, in units ofppb (parts per billion) are summarized in Table 3.7.

A two-dimensional contour plot of the concentration field is shown inFig. 3.13. It clearly shows the expected behavior, which is Gaussian. Withincreasing downstream distance, the concentration at the bank decreases. Atthe same time the plume steadily widens. Another way of looking at the datais to superimpose lateral profiles of concentration at the different x locations.This is shown in Fig. 3.14. Again, the features of the centerline decrease inconcentration and the widening of the dye plume are readily evident.

In order to extract useful information from these data, use is made of thetheoretical equation for transverse mixing, given by

C(x, y) =m

h√

4πKyxVexp

(−V y2

4Kyx

). (3.6)

In this equation, m is the mass flow rate of the dye injection, Ky is thetransverse dispersion coefficient, and h is the water depth. To make use ofthis equation, care must be taken to put all quantities into consistent units.For example, the recorded concentrations must be converted from ppb to kgm−3.

The data can be fit to this equation, in a least-squares sense, treatingboth the flow velocity and the dispersion coefficient as unknown parameters.These parameters are repeatedly adjusted until the best fit is obtained. Upondoing so, a dispersion coefficient of 0.016 m2 s−1 and a velocity of 0.38 m s−1

are obtained. Referring to Fig. 3.11, it is seen that this velocity is indeed areasonable representative value for the average near-bank velocity observednear the Dry Lake site. If the obtained parameters are then used to calculatethe theoretical concentration field (Fig. 3.15), it is seen that the results agreequite nicely with the field data (Fig. 3.13).

44

0 50 100 150 200 250 300 3500

1

2

3

4

5

6

7

8

9

10

x (m)

y (m

)

1

2

34

5

67

89

Figure 3.13: Contours (in ppb) of dye concentration for the Dry Lake near-bank steady-discharge dye study. The symbols denote the locations wherethe dye concentration was measured.

Further confirmation of the general validity of the obtained dispersioncoefficient can be found in Rutherford (1994). In Chapter 3 of his text, hesuggests that the transverse dispersion coefficient for gently irregular channelsfalls in the range of 30 to 90% of the product of water depth and shearvelocity. For the Dry Lake site, this yields a range of shear velocity from 4to 10 cm s−1, which is very much in line with the velocity results previouslydiscussed.

45

0 2 4 6 8 100

2

4

6

8

10

12

y (m)

C (

ppb)

x = 32.0 mx = 60.5 mx = 88.7 mx = 158.7 mx = 245.9 mx = 370.9 m

Figure 3.14: Lateral profiles of dye concentration at various x locations.

46

x (m) y (m) C (ppb)

32 0 10.2732 1.52 8.5132 3.05 1.09

60.5 0 7.5760.5 1.52 6.8960.5 3.05 2.5060.5 4.57 0.1188.7 0 6.1088.7 1.52 6.2588.7 3.05 3.3988.7 4.57 1.48158.7 0 5.33158.7 1.52 5.56158.7 3.05 3.84158.7 4.57 1.93245.9 0 3.71245.9 1.52 3.46245.9 3.05 2.65245.9 4.57 2.36245.9 9.14 0.06370.9 0 3.05370.9 1.52 3.05370.9 3.05 2.65370.9 4.57 2.26370.9 9.14 0.74

Table 3.7: Rhodamine dye concentrations observed during steady-input,bank-discharge dye study at Dry Lake.

47

0 50 100 150 200 250 300 3500

1

2

3

4

5

6

7

8

9

10

x (m)

y (m

)

1

2

3

45

67

89

Figure 3.15: Theoretical predictions for dye concentration at the Dry Lakesite, using the best-fit values for flow velocity and dispersion coefficient.

48

Chapter 4

Laboratory Study

Upon returning from the field, the goal of the laboratory study was to de-termine whether or not elevated levels (above ambient) of turbulence andturbidity have a measurable biological impact on rearing coho salmon. If so,then some level of regulation of boating traffic, during sensitive times of theyear, would be justified.

In reducing this to an idealized laboratory study, some compromises hadto be made. These compromises reflect the competing needs of the reproduc-tion of the field environment and the practical and economic considerationsof the laboratory environment. In the end, two different experimental trialswere devised, so as to isolate the separate effects of fluid turbulence and sus-pended sediment loading. In the first trial, coho salmon fry were exposed torepeated, brief episodes of turbulent flow. The repetition and relatively briefloading were intended to roughly reproduce the fact that, in the field, setsof breaking waves impact the river banks intermittently throughout the day.In the second trial, coho salmon fry were exposed to similarly repeated andbrief episodes of high turbidity, intended to reproduce the field observationsof sediment sloughing off of the banks.

4.1 Experimental Facilities

In order to discern a measurable trend, if any, in impact, it was decided thata minimum of three treatments (turbulence / turbidity levels) were needed.For statistical reasons, it was further decided that the experiments shouldbe conducted in triplicate. The three treatments, plus a control, therefore

49

required a total of 12 experimental tanks.These tanks were each nominally 91 cm in length, 45 cm in width, and

filled with water to a depth of 38 cm. The general idea of the arrangementof the tanks is shown in Fig. 4.1. The tanks were encased in 1.25 cm thickblueboard insulation and had lids to further insulate the tanks and to securejumping salmon. The 12 tanks were on a common sterilization and filtrationloop. The biofilter media, having a high surface area, are colonized by bac-terium that metabolizes nitrogenous wastes. The sterilization comes fromultraviolet light.

4.1.1 Turbulence Trial

Creating a spatially uniform level of turbulence in a flow is a rather difficulttask. Even in the simple case of two-dimensional open channel flow over a flatbottom, the distributions of mean flow and turbulent quantities are variable.Regarding the turbulent quantities, this is due to the fact that the sourceof production of the turbulent kinetic energy is the shear at the boundary.Thus, the layers of fluid closer to the boundary experience a higher level ofturbulence than those farther away.

One well-documented way of generating turbulence in a closed fluid do-main is through the oscillation of a ‘grid’ of rods. In this case, the turbulencecharacteristics of the flow will be functions of the rod diameter, the spacingbetween the rods, and the stroke and frequency of the grid motion.

Another well-documented flow where the turbulence characteristics arewell known is a ‘jet flow.’ In this situation, a high-velocity jet of waterdischarges from a round orifice into an ambient receiving water body. Asshown, in idealized form in Fig. 4.2, there is an initial ‘zone of establishment,’followed by a ‘zone of established flow.’ Both zones have been very wellstudied in the past, with the consequence that the spatial variation of theturbulence characteristics, as functions of the jet velocity and diameter, areknown.

Motivated by this, it was decided to set the turbulence levels in the tanksby using, in each tank, an array of four nozzles. As shown in Fig. 4.3,an intake was positioned at the far right end of the tank. The intake wasfitted with a screened inlet, in order to prevent the entrainment of fish intothe system plumbing. Water then passed through, in turn, a 1 horsepowercentrifugal pump, a valve, and a flow meter. After passing through theflow meter, the plumbing returned to the far left end of the tank, where it

50

branched into four lines. The four lines each terminated in a 1 cm diameternozzle that was mounted flush into a ‘false wall.’

The result of this configuration is a set of four overlapping jet flows.Both the mean and turbulent velocities are highest close to the jet origins.With both increasing streamwise distance (i.e. along the jet axis, away fromthe jet origin) and lateral distance (i.e. perpendicularly away from the jetaxis) these quantities decrease. For the purposes of these experiments, thespatial variability in the turbulent flow field was removed by considering avolume-averaged value. While, in principle, it is possible for an individualfish to ‘hide’ in a low-turbulence region of the tank, informal observationssuggested that this did not happen. Instead, the jet flows were strong enoughto repeatedly entrain the fish into the high-turbulence region of the tank.

4.1.2 Turbidity Trial

For the turbidity trial, which followed the turbulence trial, it was found thatsome modification to the tanks was required. The key issue here was theability of the apparatus to keep in complete suspension the sediment in thetank. Given the slightly cohesive nature of the sediment used to set theturbidity levels, it was found that the the horizontal jets were incapable ofdoing so.

In their place, a single 2.5 cm diameter pipe was oriented vertically suchthat it terminated at the bottom of the center of the tank. The pipe wasthen capped off and eight evenly-spaced holes were then drilled around thecircumference of the cap. This configuration resulted in eight jets, much likethe spokes of a wheel, parallel to and just above the tank bottom. In addition,a ninth hole, pointing downwards, was drilled in the end of the cap. Initialtrials revealed that this configuration was far superior to the horizontal jetsin terms of fully scouring the tank bottom each time the pumps were turnedon.

4.2 Experimental Procedures

4.2.1 General Details

The general experimental procedures were similar for the two sets of trials. Ineach case, the length of the trial was three weeks. To reproduce an episodic

51

loading environment, electronic timers were used to cycle the pumps on andoff. The pattern was such that a pump would come ‘on’ for 5 minutes andthen shut ‘off’ for 25 minutes. This pattern was repeated 8 times per day.

Fish were obtained from a hatchery in New York. The feed that wasused during the experiment was similar to the feed used at the hatchery andwas a sinking feed. A total of 50 fish were placed in each tank at the startof each trial. In order to assess impact, growth rate was adopted as themeasure. To facilitate this, the fish were batch weighed at the beginning ofeach experiment (Fig. 4.4). Baking soda was used to anesthetize the fish forweighing. The concentration was adjusted to the amount of water in the trayin order to achieve rapid immobilization.

Throughout the course of the three week trial, mortalities, which werevery few in number, were removed on a daily basis. At the conclusion of thethree week trial, the fish were again batch weighed. Based upon these weightsand the number of fish, a per-fish weight both before and after the trial couldbe determined. A comparison of the before and after average weight wouldthen indicate the average growth rate of the fish in an individual tank.

4.2.2 Turbulence Trials

In the turbulence trials, there were three control (C) tanks, three low treat-ment (L) tanks, three medium treatment (M) tanks, and three high treatment(H) tanks. A summary of the circulation loop flow rates and jet source ve-locities for the three treatment levels is given in Table 4.1. Also given in thetable is an estimate of the rate of energy dissipation in the tank. This isdetermined from the fact that a sharp inlet into a tank has a loss coefficientof KL = 1. From this, the head loss is determined from hL = KLV 2

0 /(2g).Finally, the power, or rate of energy loss is given by P = γhLQ. In the table,the power is normalized by the mass of water, yielding units of Watts per kg.

Treatment Level Q (liters sec−1) V0 (m s−1) P/m (W kg−1)

L 0.95 3.01 0.028M 1.6 5.03 0.13H 2.2 7.00 0.35

Table 4.1: Circulation loop volumetric flow rates, jet source velocities, andenergy dissipation levels for the three treatment levels in the turbulence trials.

52

To compare these dissipation values with the streamflow, recall that theestimates in Chapter 3 indicated shear velocities on the order of 5-8 cm s−1.Based upon the formal definition of dissipation, in units of W kg−1

Φ =∂V

∂y

(µ

∂V

∂y− ρu′v′

), (4.1)

and turbulent boundary layer solutions for open channel flow, one can showthat the cross-sectionally averaged dissipation is

Φavg =u2∗V m

h. (4.2)

In the above equations, µ is the absolute viscosity of water and V m denotesthe depth averaged value of the mean velocity.

The present data therefore yield values of dissipation around 0.002−0.005W kg−1. It must be noted, however, that these values underestimate therate of dissipation in the vicinity of the banks and river bed. From (4.1),it is clear that dissipation is greater near the boundaries, where the velocitygradients are greatest. It can be shown that the 10% of the cross sectionalarea adjacent to the boundary contains about 80% of the dissipation. Thus,with this factor of 8 increase, it is reasonable to adopt the range of 0.016−0.04W kg−1 as being a ‘baseline’ value for the dissipation in the near-bank areadue to streamflow. Thus, the low treatment level is seen to be comparableto the streamflow and the other treatments represent elevated levels.

4.2.3 Turbidity Trials

In the turbidity trial, as was previously stated, some modifications needed tobe made in order to keep the sediment wholly in suspension while the pumpswere turned on. In order to isolate the effects of the turbidity from the turbu-lence in this trial, the recirculating flows in all three treatment tanks were setto the same level. To further isolate the turbidity effects, the control tankswere equipped with plumbing and pumps identical to the treatment tanks.Put another way, the turbulence levels in all tanks were equal, ensuring thatany observed systematic differences in growth rates among the tanks wouldbe attributable to variations in turbidity levels. Given the need for pumps inthe control tanks, along with the constraint on the number of pumps avail-able, it was only possible to conduct this trial in duplicate, rather that intriplicate, as was done in the turbulence trial.

53

Treatment Level Turbidity (NTU)

L 75M 150H 250

Table 4.2: Target turbidity levels for the three treatments.

The target turbidity levels for the three treatments (L,M,H) are givenin Table 4.2. These levels were set using sediment brought back from theChilkat River following the 2003 field experiment. Levels were measured us-ing the OBS described in Chapter 3. As with the turbulence trial, the pumpsoperated in an intermittent fashion. During the ‘on’ periods, the sedimentwas completely suspended from the tank bottom and evenly distributed bythe agitation in the tank. During the ‘off’ periods, the sediment would slowlysettle out, such that the turbidity in the tank resembled an exponential-typedecay. Over a period of about 15 minutes, the water would return to a fairlyclear state.

Due to the master circulation loop that circulates water from all tanksthrough sterilization and filtration elements, the prospect of ‘losing’ sedimentfrom the treatment tanks arose. Initial efforts to prevent this loss exploredthe possibility of shutting down the main loop during the portion of the daythat the pumps were operational. This ultimately proved unsatisfactory froma water quality point of view.

As a result, some sediment loss was inevitable and was mitigated byadding additional sediment every two days. The amount of sediment lostduring a single ‘on’ pump cycle is easily estimated through consideration ofthe conservation of mass equation. Given that the inflow returning from thefiltration is sediment free, the equation for the concentration in a tank isgiven by

VdC

dt= −QC,

where V is the tank volume and Q is the through-flow. Solving this yields

C(t) = C0e−Q

Vt,

where C0 is some initial concentration. Given a tank volume of 150 litersand a flowrate of two liters per minute, it is found that, during a single five

54

minute ‘on’ cycle, approximately 7% of the sediment flows out of the tankand is trapped by the filtration system. The replacement of sediment everytwo days helped to keep the sediment levels at their target values.

55

Figure 4.1: Photograph of the experimental arrangement of tanks.

56

Figure 4.2: Schematic sketch of an axisymmetric fluid jet.

57

P

Side ViewEnd View

Top View

Figure 4.3: Schematic close-up views of the experimental apparatus for theturbulence trials.

58

Figure 4.4: Typical photograph of (sedated) fish being batch weighed.

59

4.3 Experimental Results

4.3.1 Turbulence Trials

The data from the turbulence trial is summarized in Table 4.3. The dataare given in terms of total batch weight, sample size, and normalized weight.These data are provided for both the start and the finish of the trial. In ad-dition, the average weight gain over the three week trial is provided. Finally,the ensemble averaged value (across the triplicate tanks) is provided.

60

C1

C2

C3

L1

L2

L3

M1

M2

M3

H1

H2

H3

Sta

rtD

ata

Tot

alW

eigh

t(g

)86

.48

88.2

888

.74

98.4

105.

793

.98

95.4

098

.12

92.9

086

.94

91.9

293

.38

Sam

ple

Siz

e50

5050

5054

4850

5050

5050

50A

vera

geW

eigh

t(g

)1.

731.

771.

771.

971.

961.

961.

911.

961.

861.

741.

841.

87

End

Data

Tot

alW

eigh

t(g

)91

.60

92.3

994

.63

97.3

113.

5289

.96

96.6

993

.45

94.3

879

.28

94.3

996

.20

Sam

ple

Siz

e50

4650

4954

4550

4648

4450

49A

vera

geW

eigh

t(g

)1.

832.

011.

891.

992.

102.

001.

9320

31.

971.

801.

891.

96

Avera

ge

Gain

(g)

0.10

240.

2429

0.11

790.

0177

0.14

480.

0411

0.02

570.

0691

0.10

830.

0630

0.04

940.

0957

Ense

mble

Avera

ge

Gain

(g)

0.15

440.

0679

0.06

770.

0694

Tab

le4.

3:Sum

mar

yof

wei

ght

dat

afo

rth

etu

rbule

nce

tria

l.

61

4.3.2 Turbidity Trials

The data from the turbidity trial is summarized in Table 4.4. The data aregiven in terms of total batch weight, sample size, and normalized weight.These data are provided for both the start and the finish of the trial. In ad-dition, the average weight gain over the three week trial is provided. Finally,the ensemble averaged value (across the duplicate tanks) is provided.

62

C1

C2

L1

L2

M1

M2

H1

H2

Sta

rtD

ata

Tot

alW

eigh

t(g

)32

4.9

299.

826

9.8

299.

628

6.9

290

249.

525

2.1

Sam

ple

Siz

e50

4852

5050

5050

50A

vera

geW

eigh

t(g

)6.

506.

255.

195.

995.

745.

804.

995.

04

End

Data

Tot

alW

eigh

t(g

)35

5.1

357.

833

6.9

351.

334

0.5

354.

730

5.1

300.

1Sam

ple

Siz

e50

4852

5050

5050

50A

vera

geW

eigh

t(g

)7.

107.

456.

487.

036.

817.

096.

106.

00

Avera

ge

Gain

(g)

0.60

1.21

1.29

1.03

1.07

1.29

1.11

0.96

Ense

mble

Avera

ge

Gain

(g)

0.91

1.16

1.18

1.04

Tab

le4.

4:Sum

mar

yof

wei

ghtdat

afo

rth

etu

rbid

ity

tria

l.

63

Chapter 5

Discussion and Conclusions

A near-bank hydraulic study of several sites on the Chilkat River was con-ducted during the summer of 2003. Measurements of mean velocity, turbulentvelocity, and turbidity were made with an acoustic doppler velocimeter andan optical backscatter sensor. The goal of these measurements was to estab-lish a ‘baseline’ of the typical hydraulic climate associated with the ambientflow of the river.

Subsequent to the field study, a laboratory study was conducted in orderto determine the biological response of coho salmon fry to elevated levels ofturbulence and turbidity. To assess impact, growth rate was adopted as ameasure. Separate three week trials were conducted in order to assess theimpacts of turbulence and turbidity.

5.1 Conclusions

Some of the key findings of this study include:

• Ongoing erosion pin measurements in the Chilkat River (since summer2002) demonstrate steady erosion of the banks in response the theloadings on the banks. These loadings include the stream flow andthe impacts of boat wakes. For the most part, larger erosion valuesare found in larger channels, but there are some exceptions to thisgeneralization.

• Numerous estimation methods applied to the field data arrived at ashear velocity of 5 − 10 cm s−1 in the vicinity of the banks. This

64

parameter is of broad use in terms of characterizing the mean andturbulent flow fields.

• Field measurements consistently placed the ambient, or background,turbidity of the river at ∼ 100 NTUs (nephelometric turbidity units).

• Laboratory trials revealed that the presence of fluid turbulence reducedthe growth rate of salmon fry when compared to a control group ex-posed to no turbulence. However, the reduction in growth rate ap-peared not to be a function of the turbulence level in the tank.

• Laboratory trials revealed that the presence of turbidity did not seemto affect the growth rate of salmon fry. Moreover, the growth rate didnot appear to be function of the turbidity level in the tank.

5.2 Discussion

Overall, the results suggest, within the parameters of the laboratorystudy, that coho fry are relatively unaffected physically by elevatedlevels of turbidity. Recall that these parameters include a three-weekduration, episodic (eight five-minute exposures per day) loading, andapproximately five-month old fry. Note as well that the ‘high’ treatmentlevel corresponded to a turbidity 2.5 times that of the ambient stream-flow. The relative lack of observed impact conflicts somewhat with theexperiments discussed in Chapter 2.2, which demonstrated physical im-pacts at turbidity values lower than this level. In addition, some of theprevious studies documented behavioral impacts. The present studywas not able to make these types of behavioral observations.

The present results also suggest that, within the parameters of thelaboratory study, that coho fry may be somewhat affected physicallyby the presence of elevated levels of turbulence. These parameters areequal to those described above, with the exception that the turbulencetrial was conducted with approximately three-month old fry. This con-clusion is reached by noting that all three turbulence treatment levelsyielded a growth rate noticeably lower than the control group. Thelack of variation in growth rate between the low, medium, and highttreatment levels, however, somewhat paradoxically suggests that theimpact is independent of the actual magnitude of turbulence.

65

5.3 Recommendations

As discussed previously, the present study represents a compromise.The design of a laboratory study with any level of experimental controlover the parameters inherently filters out the large degree of complexityand variability observed in the field. The main objective of the presentstudy was to investigate, in a systematic way, the effects of periodicloading at levels above those of the ambient streamflow.

The reality on the Chilkat River is that salmon fry are mobile, notcaptive. As a result, the exact length and frequency of their exposureto loading from boat wakes are highly speculative. ADFG biologistshave documented their general use of near-bank areas as habitat, butit is difficult to estimate the exact length of time that a specific fishspends in one location. Put another way, the mobility of fry acts as amitigating factor. On the other hand, the recreational season on theriver exceeds three months, whereas the experimental trial was onlythree weeks. This potentially longer exposure period may offset someof the mitigation described above.

Given the lack of experimental observations of severe physical impact, itseems reasonable to conclude that, at least in larger channels, low-leveloperation of boats poses minimal physical impact to salmon fry. Thepresent study is unable to assess whether or not use levels on the ChilkatRiver pose any behavioral impact. In extremely small channels, suchas that leading into Sheep Canyon Lake, where the hydraulic loadingis similarly small, somewhat more caution may be warranted. As theerosion pin data in that channel reveal, there is a substantial load onthe banks that can not be attributed to streamflow and therefore pointsto boating activities. The greater ratio of boat to streamflow loading inthese very small channels may result in a discernible physical impact.

66

Bibliography

ADNR 2002 Chilkat bald eagle preserve management plan: public reviewdraft. Tech. Rep.. Alaska Department of Natural Resources.

Arruda, J., Marzolf, G. & Faulk, R. 1983 The role of suspendedsediments in the nutrition of zooplankton in turbid reservoirs. Ecology64 (5), 1225–1235.

Bash, J., Berman, C. & Bolton, S. 2001 Effects of turbidity and sus-pended solids on salmonids. Tech. Rep.. Center for Streamside Studies,University of Washington.

Berg, L. 1982 The effect of exposure to short-term pulses of suspendedsediment on the behavior of juvenile salmonids. In Proceedings of the Car-nation Creek Workshop: a ten-year review (ed. G. Hartman), pp. 177–196.Department of Fisheries and Oceans, Pacific Biological Station.

Berg, L. & Northcote, T. 1985 Changes in territorial, gill-flaring, andfeeding behavior in juvenile coho salmon following short-term pulses ofsuspended sediment. Canadian Journal of Fisheries and Aquatic Sciences42, 1410–1417.

Breitburg, D. 1988 Effects of turbidity on prey consumption by stripedbass larvae. Transactions of the American Fisheries Society 117, 72–77.

Bugliosi, E. 1985 Hydrologic reconnaissance of the Chilkat River basin,southeast Alaska. Tech. Rep. 84-618. United States Geological Survey.

Gregg, R. & Bergersen, E. 1980 Mysis relicta: effects of turbidity andturbulence on short-term survival. Transactions of the American FisheriesSociety 109, 207–212.

67

Hill, D., Beachler, M. & Johnson, P. 2002 Hydrodynamic impactsof commercial jet boating on the Chilkat River, Alaska. Tech. Rep.. ThePennsylvania State University, prepared for the Alaska Department of Fishand Game.

Killgore, K., Miller, A. & Conley, K. 1987 Effects of turbulenceon yolk-sac larvae of paddlefish. Transactions of the American FisheriesSociety 116, 670–673.

Morgan, R., Rasin, V. & Noe, L. 1983 Sediment effects on eggs andlarvae of striped bass and white perch. Transactions of the American Fish-eries Society 112, 220–224.

Morgan, R., Ulanowicz, R., Jr. V. R., Noe, L. & Gray, G. 1976Effects of shear on eggs and larvae of striped bass, Morone Saxatilis, andwhite perch, M. Americana. Transactions of the American Fisheries Soci-ety 106, 149–154.

Neitzel, D., Abernethy, C. & Richmond, M. 1998 Laboratory stud-ies of the effects of turbulence and shear stresses on turbine-passed fish.Progress report. Pacific Northwest National Laboratory, Richland, WA,99352.

Newcombe, C. & MacDonald, D. 1991 Effects of suspended sedimentson aquatic ecosystems. North American Journal of Fisheries Management11, 72–82.

Rehmann, C., Stoeckel, J. & Schneider, D. 2003 Effect of turbulenceon the mortality of zebra mussel veligers. Canadian Journal of Zoology 81,1063–1069.

Rutherford, J. 1994 River Mixing . John Wiley and Sons.

Servizi, J. & Martens, D. 1987 Some effects of suspended Fraser Riversediments on sockeye salmon (Oncorhynchus nerka). In Sockeye salmonpopulation biology and future management (ed. H. Smith, L. Margolis &C. Wood), , vol. 96, pp. 254–264. Canadian Special Publication of Fisheriesand Aquatic Sciences.

68

Sigler, J., Bjornn, T. & Everest, F. 1984 Effects of chronic turbidityon density and growth of steelheads and coho salmon. Transactions of theAmerican Fisheries Society 113, 142–150.

Sutherland, A. & Ogle, D. 1975 Effect of jet boats on salmon eggs. NewZealand Journal of Marine and Freshwater Research 9 (3), 273–282.

69